Professional Documents
Culture Documents
In this simple demo, our software is presented in the form of a series of screen-shots. It would give
you the feel of how the actual software works and looks like.
Narrative is on the bottom panel of each slide. Concepts and features are explained through two
enlightening six sigma projects (one from service industry and one from manufacturing industry).
Please press ‘Enter’ or ‘Arrow’ keys on the keyboard to browse the slides at your own pace.
Process Improvement using SPC / EPC / Six Sigma Tools
There is a pressing need for the continual improvement of manufacturing and service processes in
today’s competitive business environment.
You may follow the time-tested Six Sigma DMAIC approach (Define, Measure, Analyse, Improve,
and, Control) in any improvement initiative.
2. Identify the key input & output variables (KIVs & KOVs) that are critical to quality/schedule/cost
3. Collect the relevant data and analyse them using the SPC tools such as C&E Diagram, Pareto,
Histogram, Scatter Plot, Control Chart, etc.
5. Install an on-going control scheme (such as SPC or EPC chart) to sustain the improvement.
What is SPC ?
Statistical Process control (SPC) is a scientific and inexpensive way to prevent defects. It is an
effective check against assignable causes of process variation. You would require SPC tools for
every Six Sigma project.
Both SPC and EPC are very essential for achieving the PPM (parts per million) defect levels
expected of your Six Sigma initiatives.
In this context, it is very important to provide a statistical software to your personnel for
error-free data analysis and charting on regular basis.
Our FBM Tools – SPC/EPC/Six Sigma module is a Windows-based computer program designed
specifically for improvement projects, by seasoned Engineer-Statistician experts (alumni of the
Indian Statistical Institute).
Before getting into the details of the software, let us look at one example project from service
industry and one from manufacturing industry, to learn how to put SPC/EPC/Six Sigma tools to use.
Please see the software screens and read the narrative at the bottom panel.
Example Project (Service Industry)
In a certain company, there were chronic data entry problems in the Accounts Department. As part
of its Six Sigma initiatives, the company’s management has decided to study the data entry errors
by collecting the data for one month and analysing it. The company has designated its Manager,
Mr. Stanley John as the Project Leader.
Mr. John has created a project record by going to the menu item File Manage My Projects
Example Project (Service Industry)
Then he created a new data file by going to the menu item File New Data File.
After that, he entered the data collected in August 2011 and saved the file.
Example Project (Service Industry)
For this he went to the menu item File Manage My Workbooks / Start Data Analysis Pareto
Analysis and created a workbook record.
Example Project (Service Industry)
When he clicked on the ‘START ANALYSIS NOW’ button, the software has displayed the Pareto
Diagram.
Example Project (Service Industry)
And, when he clicked on the ‘View Table’ button, the Pareto Table was shown as above.
He has decided to present this Pareto output along with a Cause & Effect Diagram to the concerned
data entry operators for further brainstorming and improvement.
Example Project (Service Industry)
He has also decided to group the errors under six categories (variation of Six M’s) such as People,
Method, Equipment, Resources, Measurements, and, Materials.
He went to the menu item File Manage My Workbooks / Start Data Analysis Cause and Effect
Diagram and created a workbook record.
Example Project (Service Industry)
When he clicked on the ‘START ANALYSIS NOW’ button, the software has displayed the Cause &
Effect Diagram.
Example Project (Service Industry)
Also, the project leader wanted to get an estimate of the current levels of Six Sigma Metrics of the
data entry process. For this, he has decided to open a control chart workbook and use the u-chart.
To setup u-chart, the project leader has gone to the menu item File Manage My Workbooks /
Start Data Analysis Control Chart and Histogram and created a workbook record.
Example Project (Service Industry)
When he clicked on the ‘START ANALYSIS NOW’ button, the software has displayed the above
message.
To view the control chart, he has gone to the menu item Graphics Control Charts and clicked
open the menu.
Example Project (Service Industry)
To view the Six Sigma metrics, he has gone to the menu item Reports Control Chart Data
Summary and clicked open the menu as above.
After checking the required sections in the report, he hit the ‘View/Print’ button to view the report.
Example Project (Service Industry)
Having done the preliminary data analysis, it is now time for brainstorming and improvement.
Example Project (Service Industry)
Looking at the Pareto Analysis, Cause & Effect Diagram, and the Control Chart, the data entry
operators as well as the project leader agreed that the root causes of the problem were:
It was decided …
- to impart training to all data entry operators on accounting concepts (correct posting)
- to enable automatic spell-check facility of the accounting software, and also to keep dictionary
CDs at data entry work stations
- to re-write the integrated management system (IMS) work instructions in such a way that cheque
details would never be lost in the communication process, and
- to continue with the u-chart for monitoring the day-to-day error levels.
Let’s see what was the result of implementing the corrective actions:
Example Project (Service Industry)
See the improved Six Sigma Metrics for September 2011. Prior to Six Sigma initiatives, these were:
Defects per Million (DPM) : 8186.15
Sigma Quality Level : 3.90
Yield (%) : 99.18
Hope that you liked the sample project from service industry.
Now, let’s see one example project from manufacturing industry.
Example Project (Manufacturing Industry)
Project Description:
In a certain pharmaceutical company that manufactures tablets, there was frequent rejection of final
product due to off-the-spec tablet weight. As part of its Six Sigma initiatives, the company’s
management decided to study the tablet weight variations by collecting some data from the plant
and analysing it.
The company has designated its Manager, Mr. Stanley John as the Project Leader and the
Laboratory Technician Ms. Ratna Raj as team member.
Now, let’s see how this very interesting project was executed.
Example Project (Manufacturing Industry)
As the first step, the project leader has created a project record by going to the menu item File
Manage My Project Records
Example Project (Manufacturing Industry)
Then, he entered the team information by clicking on the ‘View / Edit the List of Team Members’
button.
Example Project (Manufacturing Industry)
After that, he entered the key input and output variables by clicking on the ‘View / Edit the List of
Key Input and Output Variables (KIV’s and KOV’s)’ button.
Example Project (Manufacturing Industry)
Then he created a data file and entered the data that was collected with the help of team member.
Example Project (Manufacturing Industry)
He has then looked at ‘SPC Parameters’ tab and just made one entry (Target = 1.00) and kept all
others at default values. As the basic chart selected was Xbar-S and chart type selected was
‘Conventional’, these parameters were not required. Regarding ‘EPC Parameters’ tab, he decided
to enter the details at a later stage (after analysing the data using Scatter Plot and Regression tool).
Example Project (Manufacturing Industry)
He then clicked on the ‘START ANALYSIS NOW’ button, and received the above message.
Histogram, depicting the data distribution (spread) viz-a-viz tolerance band (technical
specifications).
Example Project (Manufacturing Industry)
Frequency Table (optional add-on to Histogram), showing data distribution in tabular form.
Example Project (Manufacturing Industry)
Normal Probability Plot (NPP) is a very important visual aid for checking the normality of data (i.e.,
to examine whether the data comes from a population with Normal Distribution). If the data follows
Normal Distribution, the plotted points would form a straight line.
Example Project (Manufacturing Industry)
Run chart is a simple plot of sample averages, which gives a visual understanding of patterns and
trends in control chart data.
Example Project (Manufacturing Industry)
Traditional SPC charts are not effective when the data is highly auto-correlated (i.e., when
consecutive data points are correlated. If the bars on the Auto-Correlation Chart are shorter, it
indicates less amount of auto-correlation. In case the data is highly auto-correlated, use
Un-weighted Batch Mean (UBM) chart.
Example Project (Manufacturing Industry)
Control chart shows that the process is under statistical control. That means, there is no sporadic
(assignable) cause present. The process is stable.
Example Project (Manufacturing Industry)
Six Sigma Metrics & Probability Distribution (Normal) gives an idea about expected rejections.
Though none of the data analysed were beyond specifications, the small red zone below the lower
specification limit (LSL) indicates possibility of manufacturing out-of-spec products.
Also, the process is barely capable (Cp < 1.33) and not centered (Cpk < 1).
Example Project (Manufacturing Industry)
Now, the reports were looked at. First, the data summary.
Example Project (Manufacturing Industry)
By clicking on the ‘START ANALYSIS NOW’ button, he could se the scatter plot & regression line.
As a rule of thumb, R-square value must be at least 0.70 for the regression line to be considered as
meaningful. Mr. John looked at the R-square value. It was 0.9951 (very close to the perfect value).
So, he decided to use the equation.
Example Project (Manufacturing Industry)
Setting up an Engineering Process Control (EPC) chart for the Tablet making process:
It involved the following steps (and cues from the User Manual accompanying the SPC software):
2. Setting the parameter (g) for the selected model: Overall Process Gain, g = 0.1092
3. Setting the Smoothing Constant () for EWMA predictor: = 0.15 (‘best fit’ from recent data)
4. Setting the Adjustment Boundary Value (L) for EWMA predictor: L = 0.0041
5. Re-setting the process average at the Target value (1.00 gram): Done by engineering means.
6. Installing an EPC chart and monitoring (and adjusting) the process: For this, he went back to the
SPC software.
Mr. John has edited the original control chart workbook (EPC Parameters tab), as above.
Example Project (Manufacturing Industry)
After process re-set, fresh data were taken from 2.00 PM onwards and entered in the same data file
as above.
Example Project (Manufacturing Industry)
Then, the work book entries were modified (such as project phase changed to ‘Control’, etc) as
above.
Example Project (Manufacturing Industry)
Then hit the ‘START ANALYSIS NOW’ button, and checked the EPC chart (‘Graphics’ menu).
The thin graph in black colour is the sample mean. The thick graph in blue colour is the EWMA
predictor for process average. The next point prediction (predicted value for process average at
8.00 PM) was 1.0005 and the advice was to leave the Feed Rate (X) at its present level (i.e., 2.13).
We have discussed two practical projects so far.
Never thought that such things could be done in your organisation also !
Select your User ID from the list, enter the password, and select a login option (User / Admin).
Fraction Defective (p) sample size >= 1 Visual inspection (good/bad) of lot items
Nos. Defective (np) sample size >= 1 Visual inspection (good/bad) of lot items
Defects / Unit (u) sample size > 0 Clerical mistakes (count) in data entry
Advanced SPC Charts Supported:
Short run / Standardized Same as basic charts For job-shop (small lot production)
Slopping / Tool wear 2 - 200 data / sample Where gradual process shift is inherent
Six Sigma Process 2 - 200 data / sample For monitoring Six Sigma processes
Coefficient of variation (CV)% 2 - 200 data / sample Normally, in textile and jute processing
Moving Avg. - Moving Range 1 data / sample When data is slightly auto-correlated
Moving Avg. - Moving SD 1 data / sample When data is slightly auto-correlated
CuSum 1 - 200 data / sample To detect small shifts in process mean
EWMA 1 - 200 data / sample To detect small shifts in process mean
UBM 1 data / sample When data is highly auto-correlated
Other Tools Supported: